GNGTS 2013 - Atti del 32° Convegno Nazionale

In fact, over the areas affected by the ongoing seismic sequence few ground velocity values could be measured. The relative time series show clear atmospheric seasonal signals which are however filtered out given the length of the covered period. In the larger epicentral area we could not evidence relative ground velocity variations higher than the noise level. No medium- term transients are therefore detected. However, an interesting local signal is present along the SW front of the Pollino range. In the ascending map, we observe a ground velocity gradient of ~1.5 mm/yr across the Pollino fault. A less evident gradient is observed in the descending map. Since a dense GPS network was not available across the fault at the same time of image acquisitions, we cannot validate this specific feature using independent data. However a qualitative validation can be obtained comparing our results with the velocities obtained by Sabadini et al. , 2009 using an independents stack of ERS SAR images. These authors find a ~1.5 cm/yr vertical velocity difference in the same area, i.e along the SE slope of the Pollino range, between Frascineto and Civita, with the same sense of movement of our observations. Apart from the differences in the deformation rates, the similarity of the two results suggests that the presence of artifacts in the DEM (two different DEMs were used in the two studies) or layered atmosphere can be discarded. If we discard a tectonic origin, which seems not supported by the local scale of the signal, the only remaining possibility is that the deformation pattern is due to gravitational deformation. We have investigated the presence of long-term evidences of such movements, and have identified many clear ones: old detachment zones, slope-parallel fractures, a large mid-slope valley flanked by sliding “fault” scarps, and a general bulge of the slope. All these evidences suggest that the entire slope facing Frascineto is affected by slow, possibly episodic surface movements due to deep-seated gravitational deformation, were the dominating mechanism seems to be the lateral spreading. To check the consistency of this interpretation, we used the ascending and descending velocity maps to derive the Up and East components for the common pixels in the entire area. The considered part of the Pollino slope shows a clear eastward component of movement with respect to the stable Castrovillari plain, and this is in good agreement with the geometrical arrangements of the lateral spreading mechanism driven by the visible sliding surface. In fact, deep-seated gravitational deformation on the mountain ranges forming the footwall of normal faults has been observed at several sites in the Apennines (Moro et al. , 2012, 2009, 2007; Galadini, 2006; Salvi et al. , 2003). Our preliminary interpretation will be followed by detailed field and geophysical studies of the slope. These footwall, slow moving, gravitational deformations can be excited by earthquake ground shaking up to complete collapse, and their mapping should be an important step of the seismic hazard assessment. The COSMO-SkyMed time series acquired between October 2009 and October 2010 was processed using the SBAS algorithm. Unfortunately the short period of the time series does not allow to separate the seasonal contributions from the possible tectonic signals, and the data show no convincing evidences of the presence of deformation transients before the inception of the seismic sequence in September 2010. Ground deformation occurred before October 26th, 2012. The only SAR data acquired across the October 26, 2012, Mw=5.2 event, is a COSMO data set from an ascending orbit. Unfortunately, for this data set the strong temporal decorrelation (mainly due to the high canopy vegetation cover, and to the limits of the X-band) and the presence of some temporal gaps in the image acquisition, prevented us to obtain an accurate and continuous measure of ground velocities in the epicentral area. In fact, a gap of image acquisitions of about 1.5 months precedes the mainshock, and we could not measure any deformation transient occurred in this period. 204 GNGTS 2013 S essione 1.2